BIOLOGICAL OVERVIEW

Drosophila Presenilin was isolated on the basis of shared sequences with mammalian presenilins. Mutations in the two human presenilins, PS1 and PS2, and in another protein, the amyloid precursor protein (APP), are associated with early onset familial Alzheimer's disease (AD). Presenilins have two known functions: they affect the processing of ß-amyloid precursor protein and facilitate the activity of transmembrane receptors of the Notch family. Before discussing the role of presenilins in Notch processing, their influence on ß-amyloid protein processing will be described.

ß-Amyloid precursor protein (ß-APP) is a transmembrane protein that travels by way of the endoplasmic reticulum and Golgi to the cell surface and undergoes proteolytic processing. A set of ß-amyloid (Aß) peptides are generated from ß-APP by proteases known as the ß- and gamma-secretases. The ß-secretase cleavage occurs in the extracellular domain and the heterogeneous gamma-secretase cleavages occur in the transmembrane domain. Dominant mutations in either of two Presenilin genes appear to cause Alzheimer's disease by increasing the amount of the Aß42(43) fragment that is produced. A null allele of mouse Presenilin 1 appears selectively to reduce-gamma-secretase activity (DeStrooper, 1998). These observations indicate that presenilin either stimulates the activity of gamma-secretase, or is itself a component of gamma-secretase (Struhl, 1999 and references).

Evidence has been found of a role for Drosophila Psn in Notch processing. Notch acts as a transmembrane cell-surface receptor for intercellular signals during development. It has been proposed that signal transduction involves cleavage and transport of the Notch intracellular domain to the nucleus. Results from Drosophila and mammalian cells indicate that cleavage occurs in or near the transmembrane domain (Struhl, 1998; Schroeter, 1998; Lecourtois, 1998, and Kidd, 1998). In mammalian cells, at least one proteolytic event occurs in the extracellular domain during Notch transit to the cell surface (Logeat, 1998), and it has been suggested that ligand-binding might trigger additional extracellular proteolytic processing. Thus Notch proteins undergo proteolytic processing events that resemble the ß- and gamma-secretase cleavages of ß-APP. These parallels, as well as genetic studies of presenilin in C. elegans, indicate that the presenilins may promote proteolytic cleavage during receptor maturation or activation (Levitan, 1998)

To investigate the involvement of presenilin in proteolysis of Notch protein, an in vivo assay was used for ligand-dependent cleavage and nuclear access of the intracellular domain of Drosophila Notch. To identify Psn mutants, an examination was made of a collection of recessive-lethal mutations that map to the location identified for Psn and that cause a neurogenic phenotype in genetic mosaics (J. Jiang, C.-M. Chen and G. Struhl cited in Struhl, 1999). Two independent mutations, PS^C1 anPS^C2, contain lesions in Psn. Both alleles are predicted to cause premature termination and appear to be null alleles. To generate embryos with no Psn activity, both maternal and zygotic Psn activity were removed by generating Psn minus embryos derived from Psn minus germ cells. In both Psn minus and Notch minus embryos, clusters of neuroblasts segregate at the positions normally occupied by single neuroblasts, as revealed by Hunchback staining. Both Psn minus and Notch minus embryos also show extensive neural hyperplasia during subsequent development and die as pharate first-instar larvae lacking both dorsal and ventrical cuticle. In addition, the number of midline cells, as defined by the expression of Single-minded (Sim), is greatly reduced. Notch protein is found predominantly at the plasma membrane and at similar levels in both wild-type and Psn minus embryos. Hence, the profound developmental defects in Psn minus embryos appears to result from the absence of Notch signal-transducing activity, rather than from a marked decrease in Notch protein at the plasma membrane (Struhl, 1999).

The effect of Psn null mutation on nuclear access by the Notch intracellular domain was examined by using three Notch proteins in which the chimaeric transcription factor Gal4-VP16 (GV) was inserted in-frame into Notch just after the transmembrane domain. Nuclear access was assayed by UAS-lacZ expression. N⁺-GV3 functions like the wild-type Notch protein and the intracellular domain gains nuclear access and has signal transducing activity only in the presence of the ligand, Delta. N^ECN-GV3 contains a deletion that removes most of the extracellular domain and causes constitutive signal transducing activity and nuclear access in the absence of Delta. N^intra-GV3 lacks the extracellular and transmembrane domains and also displays ligand-independent nuclear access. The key difference between the two constitutively active forms is that N^ECN-GV3 retains the transmembrane and extracellular juxtamembrane domains, whereas N^intra-GV3 is a cytosolic protein (Struhl, 1999).

In Psn minus embryos, neither N⁺-GV3 nor N^ECN-GV3 has access to the nucleus, as indicated by the complete absence of ß-galactosidase (ß-Gal) expression. In contrast, the nuclear access of N^intra-GV3 is unaffected by the absence of presenilin activity. The N⁺-GV3 observation indicates that presenilin activity is normally required for the nuclear access of Notch intracellular domain. Furthermore, the observation that presenilin is needed for nuclear access of N^ECN-GV3, a constitutively active transmembrane form, but not for N^intra-GV3, a constitutively active cytosolic form, suggests that presenilin participates in the release of the intracellular domain from the plasma membrane. Only about 35 amino acids of the Notch extracellular juxtamembrane region remain in the N^ECN-GV3 protein. Thus, if there are specific signals required for presenilin-dependent cleavage, they are likely to be somewhere in this region or within the transmembrane domain (Struhl, 1999). Similar experiments by Y. Ye (1999) confirm these results.

Although these experiments demonstrate that presenilin is necessary for the ligand-dependent nuclear access of the intracellular domain of Notch, it is not known whether presenilin directly mediates proteolytic release of the intracellular domain or if it acts more indirectly, for example by activating a protease or mediating the protease's transit to the plasma membrane. These findings can be incorporated into a model of events involved in Notch signal transduction, in which ligand-binding activates Notch, thereby creating a substrate for presenilin-dependent release of the intracellular domain from the membrane. Although there are other possibilities, release could require the direct participation of presenilin in the proteolytic cleavage of Notch protein in or near the transmembrane domain. Presenilin may play an analogous role in the processing of ß-APP (Struhl, 1999). Experiments with mammalian Notch1 and PS1 show that the two proteins physically interact. The interaction predominantly occurs early in the secretory pathway, prior to Notch cleavage in the Golgi, because PS1 immunoprecipitation preferentially recovers the full-length Notch1 precursor. These results suggest that the genetic relationship between presenilins and the Notch signaling pathway derives from a direct physical association between these proteins in the secretory pathway (Ray, 1999).

What is the function of presenilin? Given the evidence that presenilin is required for processing of Notch and APP at transmembrane sites, there are a number of possibilities for the function of presenilin. One is that presenilin is required for the proper trafficking of Notch and APP to their protease(s), which may reside in an intracellular compartment. Another is that presenilin is required for the proper biogenesis or trafficking of the gamma-secretase, the protease thought to target APP. Another is that presenilin is an essential cofactor for gamma-secretase, and yet another is that presenilin itself is gamma-secretase. The current data cannot distinguish among these possible functions. Chan (1999) refers to Wolfe et al., (1999), who favor the idea that presenilin is itself gamma-secretase. Wolfe et al. focused on two transmembrane aspartate residues conserved in all known presenilins. Expression in cultured cells of presenilin constructs in which either of these two residues is mutated results in loss of presenilin activity as assayed by production of Aß, even though these cells express wild-type presenilin-1 endogenously. It was therefore concluded that these constructs act as dominant-negative presenilins. While wild-type presenilins are cleaved at a site in the cytosolic loop, the authors found that the aspartate mutants are not cleaved in this manner; they take this result as circumstantial evidence that this cleavage is mediated by presenilin-1 itself. They also point out that gamma-secretase has some characteristics of aspartyl proteases and speculate that the conserved aspartate residues may contribute to an active protease site in presenilin. Wolfe et al. further found that Aß can be produced by in vitro translation of an APP-derived construct in the presence of microsomes derived from wild-type cells but not when the microsomes are prepared from presenilin-1 mutant cells. It was reasoned that presenilin-1 is not required for trafficking of APP, as little or no vesicular trafficking is expected to occur in the in vitro microsomal preparations, and the authors suggest that presenilin is required in the same subcellular compartment in which gamma-secretase resides. An alternative possibility is that presenilin is required for the proper production, processing, or localization of gamma-secretase or an essential cofactor. Presenilin could also have multiple functions, including trafficking of APP, only one of which is required in microsomes. As Wolfe et al. point out, conclusive evidence that presenilin is gamma-secretase will require a purified, reconstituted system, which may be technically difficult to accomplish. A somewhat more tractable, but less conclusive, approach would be to determine whether the newly identified gamma-secretase inhibitors can bind directly to presenilin with the appropriate kinetics (Chan, 1999).

Even should the presenilins not have a role in trafficking, endocytosis appears to have an important role in the processing of APP and in the function of Notch. Inhibitors of endocytosis partially block the processing of APP by gamma-secretase in cultured cells. shibire, which encodes the Drosophila dynamin small GTPase and is required for endocytosis, appears to be required for activation of Notch upon ligand binding. Trafficking of APP and Notch could regulate accessibility to their protease(s). Hence, determining the subcellular compartments for various processing events may contribute importantly to understanding the events in Notch signaling and Aß production as well as their regulation. The normal function of the presenilins is still elusive, and many other questions remain to be answered. How do the mutations associated with AD affect presenilin function? What is the extent of overlap between the Notch and APP processing machineries? Other genes implicated in the Notch pathway might also have a role in APP processing. Are Notch and APP the only molecules that require presenilins for proteolysis? A careful analysis of the phenotypes of presenilin mutants may reveal defects not found in Notch mutants. By unraveling the connection between Notch and the presenilins, researchers may find answers to long-standing questions regarding both Notch signaling and the pathogenesis of Alzheimer's disease (Chan, 1999 and references).

GENE STRUCTURE

In two of the Psn sequences examined, an internal in-frame deletion of 42 base pairs was observed, clearly suggesting the existence of an alternatively spliced exon. This 42 bp fragment encodes 14 amino acids embedded in the homologous region to the large hydrophilic loop HL-VI between transmembrane regions VI and VII of human PS-1 and PS-2 proteins (Marfany, 1998).

Exons - 9

PROTEIN STRUCTURE

Amino Acids - 541

Structural Domains

Comparison of the predicted amino aid sequence of the longest reading frame of Drosophila Psn with that of human PS1 reveals 53% overall sequence identity. Residues predicted to compose transmembrane or membrane-associated domains (especially TM1, TM3, TM5, TM6, TM7, the hydrophobic domains at the beginning and the end of the TM6-TM7 loop domain, and the C-terminus) show a much higher sequence homology. Interesting, 16 of 20 residues mutated in human PS1 or PS2 and giving rise to human familial AD are conserved in Drosophia Psn. In addition, the cystein residue that is mutated in the C. elegans sel-12 presenilin homolog is also conserved. This suggests that these residues form an important functional or structural domain of the presenilins. There is no sequence conservation at the N-terminus or at the apex of the TM6-TM7 loop domain, consistent wih the observation that there is a lack of homology in those domains between PS1 and PS2, and between the presenilins and sel-12. There are indications that Drosophila Psn is alternatively spliced. The Drosophila protein appears to be longer than human presenilins and differences in the length of two hydrophilic regions account for the difference (Boulianne, 1997, Hong, 1997 and Marfany, 1998)

To facilitate the identification of mutations in Psn, genomic clones encompassing the Psn locus have been isolated and characterized. The Psn gene is entirely contained within 4 kb of genomic DNA and is flanked by the genes encoding lipoic acid synthase and the 50 S ribosomal protein L15. The Psn transcript is encoded by nine exons, eight of which comprise the coding sequence. Overall, the structure of the Psn gene is highly related to both the vertebrate and C. elegans presenilins, and many of the intron-exon boundaries and splice sites are conserved. The Psn gene also gives rise to at least two distinct isoforms, dpsa1 and dpsa2, that result from differential splicing of exon 7. dpsa2 differs from dpsa1 by an additional 14 amino acids that are located within the large hydrophilic loop between TM6 and TM7. This loop region is highly variable in both length and amino acid sequence between species: it remains to be determined whether dpsa and dpsb are differentially expressed or have differential functions during development (Guo, 1999).

date revised: 28 June 99

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.